Results for 2017 campaign

This series of files compile all analyses done during Chapter 1 (analyses_1, analyses_2, analyses_cartography, kriging), with a division by campaign (2014, 2016 and 2017).

To navigate on this page:

Details about softwares and files used:

To return to Chapter 1’s index page, click here.

Workspace preparation

Loading of the datasets

Here, we use data from subtidal ecosystems (see metadata files for more information)
Only stations that have been sampled both for abiotic parameters and benthic species were included.
Note that the sieving method is not the same between campaigns: 2014 and 2017 campaigns used 500µm sieves and 2016 used 1000µm sieves.
The script below includes personnal functions, refined data, parameters for each campaign and global means, sd, se.
Also includes heavy metals concentrations for campaign 2017 obtained after inverse weighting distance kriging (see data_creation script).

Selected variables are:
- depth of the station: depth
- percentage of organic matter: om
- percentage of gravel: gravel
- percentage of sand: sand
- percentage of silt: silt
- percentage of clay: clay
- concentration of arsenic: arsenic
- concentration of cadmium: cadmium
- concentration of chromium: chromium
- concentration of copper: copper
- concentration of iron: iron
- concentration of manganese: manganese
- concentration of mercury: mercury
- concentration of lead: lead
- concentration of zinc: zinc
- species richness: S
- abundance of total species individuals: N
- Shannon index: H
- Piélou evenness: J
- abundance of Bipalponephtys neotena: Bneo
- abundance of Macoma calcarea: Mcal
- abundance of Echinarachinus parma: Epar
- abundance of Nematoda: Nema

Species were selected using the IndVal calculation and SIMPER procedure in PRIMER, and correlations between parameters were checked (see below).
As there is not the same number of replicates for metal concentrations and the rest of the variables, two sets of data have been produced: one with all stations available for each analysis (full dataset), and one with stations only at BSI (reduced dataset).

Graphical explorations

1. Parameters

Here are the barlots for the values of each selected variable at each BSI and MR stations. Barplots have been created with the full datasets.

Results of univariate PermANOVAs on parameters and multivariate PermANOVA on the whole benthic community (PRIMER)

Variable Region Comments
depth
om S
gravel
sand S
silt S
clay
Bneo (500 µm) S
Mcal (500 µm) S
Epar (500 µm) S
Nema (500 µm) S
S (500 µm) S
N (500 µm)
H (500 µm)
J (500 µm)
ALL SPECIES (500 µm) S
Bneo (1 mm) S
Mcal (1 mm) S
Epar (1 mm) S
Nema (1 mm) S
S (1 mm) S
N (1 mm)
H (1 mm)
J (1 mm)
ALL SPECIES (1 mm) S

Depth

## Performing barplot for:   mean  (legend is:  depth )

Organic matter

## Performing barplot for:   mean  (legend is:  om )

Grain-size

## Using group as id variables
## Performing barplot for:   mean  (legend is:  gravel )

## Performing barplot for:   mean  (legend is:  sand )

## Performing barplot for:   mean  (legend is:  silt )

## Performing barplot for:   mean  (legend is:  clay )

Heavy-metals

This section is not relevant here, as we only have data for BSI.

Species abundances

## Using group as id variables
## Performing barplot for:   mean  (legend is:  Bneo )

## Performing barplot for:   mean  (legend is:  Mcal )

## Performing barplot for:   mean  (legend is:  Epar )

## Performing barplot for:   mean  (legend is:  Nema )

Diversity indices

## Using group as id variables
## Performing barplot for:   mean  (legend is:  S )

## Performing barplot for:   mean  (legend is:  N )

## Performing barplot for:   mean  (legend is:  H )

## Performing barplot for:   mean  (legend is:  J )

2. Phylum frequencies

Phylum frequencies have been calculated with the full datasets.

3. Species estimation curves

Accumulation and rarefaction curves have been calculated with the full datasets.

4. Principal Component Analysis

All abiotic variables at BSI

Variables have been scaled by mean and standard-deviation prior to analysis.

Parameters

Variables have been scaled by mean and standard-deviation prior to analysis.

Metals at BSI

Variables have been scaled by mean and standard-deviation prior to analysis.

5. Non-metric Multidimensional Scaling

All abiotic parameters at BSI

Variables have been scaled by mean and standard-deviation prior to analysis.

Parameters

Variables have been scaled by mean and standard-deviation prior to analysis.

Metals at BSI

Variables have been scaled by mean and standard-deviation prior to analysis.

Species

Stations with no species are deleted from this analysis.

6. Hierarchical Agglomerative Clustering

All abiotic variables at BSI

Variables have been scaled by mean and standard-deviation prior to analysis.

## 127 128 129 130 131 132 134 135 136 137 138 139 140 141 142 143 144 145 
##   1   1   1   2   1   1   1   1   1   1   3   3   1   1   1   1   1   1 
## 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 
##   1   1   1   1   1   1   2   2   2   2   2   2   2   2   2   2   2   2 
## 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 
##   2   2   3   3   4   3   3   3   3   4   4   4   4   4   4   4   4   4 
## 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 
##   4   3   3   4   4   3   1   2   2   2   2   1   1   3   2   2   3   1 
## 200 201 202 203 204 205 206 207 208 209 211 212 214 215 216 217 218 219 
##   2   2   3   3   4   4   4   4   4   1   4   4   4   1   2   3   1   2 
## 220 221 222 223 224 225 226 228 229 230 231 232 233 234 235 236 237 238 
##   3   4   4   4   4   4   4   1   2   3   1   3   3   3   4   4   3   2 
## 239 240 241 
##   2   1   2
## Warning in rm(hac_spec17, groups1): objet 'hac_spec17' introuvable

Parameters

Variables have been scaled by mean and standard-deviation prior to analysis.

## 127 128 129 130 131 132 134 135 136 137 138 139 140 141 142 143 144 145 
##   1   2   3   4   2   2   2   2   2   2   3   1   1   2   2   2   3   2 
## 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 
##   2   2   2   2   2   2   1   1   1   1   1   1   1   1   4   1   1   1 
## 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 
##   1   1   4   4   1   4   4   4   4   1   1   1   1   1   1   1   1   1 
## 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 
##   1   3   4   1   1   4   1   1   1   1   1   1   2   4   1   1   4   2 
## 200 201 202 203 204 205 206 207 208 209 211 212 214 215 216 217 218 219 
##   1   1   3   4   1   1   1   1   1   3   4   4   4   3   1   4   1   4 
## 220 221 222 223 224 225 226 228 229 230 231 232 233 234 235 236 237 238 
##   1   1   4   4   4   4   1   1   1   4   1   4   4   4   1   1   3   1 
## 239 240 241 243 246 248 249 251 252 254 255 257 258 260 261 264 267 270 
##   1   2   1   3   4   4   3   4   3   4   4   4   4   4   4   4   4   4

Metals at BSI

Variables have been scaled by mean and standard-deviation prior to analysis.

## 127 128 129 130 131 132 134 135 136 137 138 139 140 141 142 143 144 145 
##   1   1   1   1   2   2   1   1   1   1   1   3   1   1   1   1   1   1 
## 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 
##   1   1   1   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2 
## 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 
##   2   2   3   3   3   3   3   3   3   3   3   3   3   3   3   3   3   3 
## 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 
##   3   3   3   3   3   2   2   2   2   2   2   2   3   3   2   2   3   2 
## 200 201 202 203 204 205 206 207 208 209 211 212 214 215 216 217 218 219 
##   2   2   3   3   3   3   3   3   3   2   3   3   3   2   2   3   2   2 
## 220 221 222 223 224 225 226 228 229 230 231 232 233 234 235 236 237 238 
##   3   3   3   3   3   3   3   1   3   3   1   3   3   3   3   3   3   2 
## 239 240 241 
##   2   2   2

Species

## 127 128 129 130 131 132 134 135 136 137 138 139 140 141 142 143 144 145 
##   1   2   2   1   2   2   2   2   2   2   2   1   2   2   2   2   2   2 
## 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 
##   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   1 
## 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 
##   2   2   1   1   2   1   1   1   1   1   2   1   2   2   2   2   2   2 
## 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 
##   2   2   1   2   2   1   1   2   1   2   2   1   1   1   2   2   1   2 
## 200 201 202 203 204 205 206 207 208 209 211 212 214 215 216 217 218 219 
##   2   2   1   2   1   2   1   2   2   1   2   1   2   1   1   1   2   1 
## 220 221 222 223 224 225 226 228 229 230 231 232 233 234 235 236 237 238 
##   1   2   2   2   1   2   2   1   2   1   2   1   1   1   1   2   1   1 
## 239 240 241 243 246 248 249 251 252 254 255 257 258 260 261 264 267 270 
##   2   2   2   1   1   1   1   1   1   1   1   1   1   1   1   1   1   1

Analyses

1. Correlations between parameters

Correlations between habitat and metal variables have been done with the reduced datasets only.

Parameters

Correlation coefficients between habitat parameters (BSI and MR stations)
  depth om gravel sand silt clay
depth 1 0.311 -0.061 -0.142 0.116 -0.117
om 0.311 1 -0.361 -0.721 0.741 0.037
gravel -0.061 -0.361 1 0.105 -0.407 0.12
sand -0.142 -0.721 0.105 1 -0.798 -0.386
silt 0.116 0.741 -0.407 -0.798 1 0.087
clay -0.117 0.037 0.12 -0.386 0.087 1

Metals

Correlation coefficients between habitat parameters and metals concentrations (BSI stations only)
  arsenic cadmium chromium copper iron manganese mercury lead zinc
arsenic 1 0.709 0.745 0.812 0.674 0.792 0.695 0.9 0.825
cadmium 0.709 1 0.813 0.698 0.576 0.751 0.649 0.819 0.821
chromium 0.745 0.813 1 0.856 0.824 0.916 0.623 0.808 0.891
copper 0.812 0.698 0.856 1 0.785 0.812 0.638 0.888 0.961
iron 0.674 0.576 0.824 0.785 1 0.863 0.356 0.614 0.751
manganese 0.792 0.751 0.916 0.812 0.863 1 0.544 0.763 0.807
mercury 0.695 0.649 0.623 0.638 0.356 0.544 1 0.816 0.701
lead 0.9 0.819 0.808 0.888 0.614 0.763 0.816 1 0.921
zinc 0.825 0.821 0.891 0.961 0.751 0.807 0.701 0.921 1

Species

Correlation coefficients between diversity indices
  Bneo Mcal Epar Nema S N H J
Bneo 1 0.374 -0.382 -0.276 0.467 0.518 0.215 -0.206
Mcal 0.374 1 -0.212 0.048 0.393 0.339 0.315 -0.023
Epar -0.382 -0.212 1 0.3 -0.145 -0.05 -0.172 -0.139
Nema -0.276 0.048 0.3 1 0.166 0.231 0.051 -0.12
S 0.467 0.393 -0.145 0.166 1 0.576 0.722 -0.013
N 0.518 0.339 -0.05 0.231 0.576 1 0.004 -0.64
H 0.215 0.315 -0.172 0.051 0.722 0.004 1 0.615
J -0.206 -0.023 -0.139 -0.12 -0.013 -0.64 0.615 1

2. IndVal and SIMPER

These analyses allowed to select 4 species as dependant variables for the regressions. We used results from PRIMER to justify further their choice. Full datasets were used.

##                         cluster indicator_value probability
## bipalponephtys_neotena        1          0.7637       0.001
## macoma_calcarea               1          0.7560       0.001
## eudorellopsis_integra         1          0.5569       0.006
## protomedeia_grandimana        1          0.4054       0.036
## goniada_maculata              1          0.3694       0.024
## echinarachnius_parma          2          0.7211       0.001
## nematoda                      2          0.6006       0.001
## crenella_decussata            2          0.5881       0.001
## ecrobia_truncata              2          0.3333       0.001
## phoxocephalus_holbolli        2          0.2927       0.050
## ameritella_agilis             2          0.2791       0.011
## mesodesma_arctatum            2          0.2667       0.001
## orchomenella_minuta           2          0.2145       0.013
## mysella_planulata             2          0.2000       0.002
## hiatella_arctica              2          0.1909       0.004
## parvicardium_pinnulatum       2          0.1792       0.018
## solariella_sp                 2          0.1708       0.033
## ophelia_limacina              2          0.1663       0.039
## anthozoa                      2          0.1276       0.022
## 
## Sum of probabilities                 =  83.633 
## 
## Sum of Indicator Values              =  15.21 
## 
## Sum of Significant Indicator Values  =  7.18 
## 
## Number of Significant Indicators     =  19 
## 
## Significant Indicator Distribution
## 
##  1  2 
##  5 14

Quitting from lines 663-673 (Chap1_article_2017.Rmd) Error in cumsum <= 0.8 : comparaison (4) possible seulement pour les types liste et atomique

3. Univariate regressions

Sand variables (Csand, Msand, Fsand) and mud variables (silt, clay) were merged to reduced the problem of model overfitting. Regressions have been splitted to include only sediment parameters or metals concentrations at a time.

All abiotic parameters at BSI

i) Simple regressions

These analyses have been do to explore the relationships between variables. As it is a huge number of results to interpret, only multiple regressions will be included in the article (see below).

Adjusted R-squared of simple regressions with all variables at BSI
  depth om gravel sand silt clay arsenic cadmium chromium copper iron manganese mercury lead zinc
Bneo 0.02896 -0.004853 -0.001284 0.005544 0.03127 -0.003883 0.02658 0.05792 0.03385 0.05284 -0.007023 -0.007612 0.01923 0.06944 0.09318
Mcal -0.005863 0.02576 -0.001422 0.02124 0.006305 -0.00896 0.04531 0.06456 0.03261 0.004907 -0.007508 0.03006 0.06613 0.06284 0.01137
Epar 0.01171 0.0884 -0.008427 0.08001 0.05474 -0.006003 0.01663 0.03255 0.0243 0.02964 0.002643 0.02949 -0.005137 0.02621 0.03627
Nema -0.006172 0.08653 0.08026 0.02622 0.07877 -0.006292 -0.009159 0.002577 -0.005425 0.005019 -0.003006 0.003156 0.0006913 0.00456 0.001428
S 0.1368 -0.009104 -0.007717 -0.008807 -0.008576 -0.00739 0.01662 0.1027 0.118 0.05078 0.1085 0.07493 0.002283 0.03525 0.06309
N 0.02846 0.02306 -0.003185 0.008909 0.004304 0.0004222 -0.006692 -0.004822 -0.008279 -0.003719 -0.001477 -0.002075 -0.008671 0.005467 0.005936
H 0.3053 0.02337 -0.008658 0.00157 -0.005121 -0.0008101 0.05214 0.1186 0.1572 0.09365 0.06617 0.0456 0.00858 0.086 0.1237
J 0.1429 0.02409 -0.00736 1.223e-05 -0.007634 0.01008 0.06716 0.02249 0.06068 0.07452 -0.00309 -0.0007157 0.004781 0.07601 0.08759
p-values of simple regressions with all variables at BSI (continued below)
  depth om gravel sand silt clay arsenic cadmium chromium copper iron manganese
Bneo 0.04091 0.495 0.3561 0.2067 0.03515 0.4501 0.04787 0.006292 0.02968 0.008709 0.6304 0.6818
Mcal 0.5504 0.05056 0.3603 0.06842 0.1953 0.8792 0.01413 0.004114 0.03219 0.2169 0.6719 0.03805
Epar 0.132 0.0008946 0.7768 0.001532 0.00771 0.559 0.09367 0.03233 0.05573 0.03912 0.2583 0.03952
Nema 0.5697 0.001009 0.001508 0.04904 0.001659 0.5775 0.9675 0.2596 0.5251 0.2151 0.4147 0.2481
S 3.837e-05 0.9306 0.6922 0.8425 0.7997 0.6613 0.09373 0.0003555 0.0001322 0.009938 0.000244 0.002121
N 0.04227 0.06057 0.4216 0.1613 0.2271 0.3086 0.6052 0.4935 0.7563 0.4431 0.3621 0.3815
H 1.944e-10 0.0593 0.8138 0.2812 0.5087 0.342 0.009112 0.0001268 9.806e-06 0.0006383 0.003713 0.01387
J 2.554e-05 0.05654 0.6586 0.3192 0.684 0.1482 0.003486 0.0629 0.005274 0.002177 0.4179 0.3393
  mercury lead zinc
Bneo 0.0784 0.003012 0.000658
Mcal 0.003723 0.004594 0.1352
Epar 0.5096 0.04909 0.02535
Nema 0.3019 0.2227 0.2844
S 0.2657 0.0271 0.004519
N 0.8161 0.2079 0.2008
H 0.1652 0.001044 9.084e-05
J 0.219 0.001979 0.0009426
ii) Multiple regressions

ZIP models are “computationally” singular, so they have not been computed here.

This table is summarizing the significative relationships with all variables as predictors, obtained with the AIC-reduced models from the multiple regressions below (ZIP not included)

Predictor Bneo Mcal Epar Nema S N H J
depth - - + + - + +
om + - - - +
gravel - - - +
sand - + +
silt
clay - + +
arsenic + - + -
cadmium - - - - - - - +
chromium + - + + -
copper - - + -
iron - - - -
manganese - + - - - +
mercury - + +
lead + - - + + -
zinc + - - +
McFadden Pseudo-R2 0.31 0.30 0.44 0.63
Adjusted-R2 0.25 0.09 0.40 0.30
Abundance of B. neotena
## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.31
Fitting generalized (poisson/log) linear model: Bneo ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 2.86 0.2605 10.98 4.999e-28 * * *
depth -0.01449 0.001937 -7.482 7.335e-14 * * *
om 0.3788 0.04618 8.203 2.341e-16 * * *
gravel -1.228 0.3515 -3.495 0.0004743 * * *
sand -0.09153 0.2125 -0.4308 0.6666
clay -3.001 0.4623 -6.491 8.551e-11 * * *
arsenic 0.1742 0.02991 5.825 5.699e-09 * * *
cadmium -19.48 2.387 -8.16 3.344e-16 * * *
chromium 0.0194 0.00525 3.696 0.0002189 * * *
copper -0.1212 0.0256 -4.736 2.179e-06 * * *
iron -2.74e-05 5.594e-06 -4.898 9.672e-07 * * *
manganese -0.001179 0.0001475 -7.995 1.297e-15 * * *
mercury -0.7621 2.702 -0.2821 0.7779
lead 0.1482 0.05692 2.603 0.009241 * *
zinc 0.07154 0.01024 6.984 2.878e-12 * * *
Variance Inflation Factors
  depth om gravel sand clay arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 1.53 2.04 1.04 2.02 1.09 2.24 3.39 4.03 5.86 2.94 3.11 1.62 4.9 8.6 

## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.31
Fitting generalized (poisson/log) linear model: Bneo ~ depth + om + gravel + clay + arsenic + cadmium + chromium + copper + iron + manganese + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 2.776 0.1789 15.52 2.539e-54 * * *
depth -0.01435 0.001908 -7.519 5.514e-14 * * *
om 0.3913 0.0357 10.96 6.035e-28 * * *
gravel -1.213 0.3499 -3.466 0.0005273 * * *
clay -2.921 0.4192 -6.968 3.209e-12 * * *
arsenic 0.176 0.02926 6.015 1.796e-09 * * *
cadmium -19.72 2.303 -8.563 1.096e-17 * * *
chromium 0.01936 0.00524 3.695 0.0002197 * * *
copper -0.1237 0.02479 -4.989 6.071e-07 * * *
iron -2.783e-05 5.404e-06 -5.15 2.599e-07 * * *
manganese -0.00118 0.0001468 -8.036 9.244e-16 * * *
lead 0.1461 0.04961 2.945 0.003232 * *
zinc 0.07303 0.009673 7.55 4.356e-14 * * *
Variance Inflation Factors
  depth om gravel clay arsenic cadmium chromium copper iron manganese lead zinc
VIF 1.51 1.58 1.03 1 2.2 3.27 4.02 5.68 2.85 3.1 4.27 8.11 

## Analysis of Deviance Table
## 
## Model 1: Bneo ~ depth + om + gravel + sand + clay + arsenic + cadmium + 
##     chromium + copper + iron + manganese + mercury + lead + zinc
## Model 2: Bneo ~ depth + om + gravel + clay + arsenic + cadmium + chromium + 
##     copper + iron + manganese + lead + zinc
##   Resid. Df Resid. Dev Df Deviance
## 1        95     2443.9            
## 2        97     2444.1 -2 -0.19436
Abundance of M. calcarea
## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.3
Fitting generalized (poisson/log) linear model: Mcal ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 4.588 0.4109 11.17 5.907e-29 * * *
depth -0.009845 0.003206 -3.07 0.002137 * *
om -0.3777 0.09957 -3.793 0.0001487 * * *
gravel -2.593 0.7163 -3.62 0.000294 * * *
sand -1.36 0.3342 -4.068 4.738e-05 * * *
clay -0.01686 0.5468 -0.03083 0.9754
arsenic -0.231 0.09091 -2.541 0.01105 *
cadmium -26.48 5.41 -4.895 9.853e-07 * * *
chromium -0.02734 0.007081 -3.861 0.0001131 * * *
copper -0.1341 0.05442 -2.464 0.01373 *
iron -1.315e-06 3.155e-06 -0.4169 0.6767
manganese 0.0009874 0.0002701 3.656 0.0002564 * * *
mercury -22.68 7.502 -3.023 0.002505 * *
lead -0.3859 0.1199 -3.218 0.001289 * *
zinc 0.122 0.01902 6.414 1.413e-10 * * *
Variance Inflation Factors
  depth om gravel sand clay arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 1.36 1.71 1.06 1.8 1.27 2.41 3.63 2.6 6.33 1.39 2.54 1.92 4.83 7.48 

## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.3
Fitting generalized (poisson/log) linear model: Mcal ~ depth + om + gravel + sand + arsenic + cadmium + chromium + copper + manganese + mercury + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 4.561 0.3822 11.93 7.885e-33 * * *
depth -0.009956 0.003185 -3.126 0.001775 * *
om -0.3808 0.09563 -3.982 6.835e-05 * * *
gravel -2.581 0.7085 -3.644 0.0002689 * * *
sand -1.381 0.2779 -4.97 6.709e-07 * * *
arsenic -0.2346 0.09011 -2.604 0.00922 * *
cadmium -26.14 5.322 -4.91 9.093e-07 * * *
chromium -0.02785 0.006973 -3.994 6.489e-05 * * *
copper -0.1348 0.05407 -2.493 0.01266 *
manganese 0.0009644 0.000266 3.626 0.0002879 * * *
mercury -22.1 7.267 -3.041 0.002355 * *
lead -0.3849 0.1193 -3.225 0.001258 * *
zinc 0.1217 0.01886 6.455 1.082e-10 * * *
Variance Inflation Factors
  depth om gravel sand arsenic cadmium chromium copper manganese mercury lead zinc
VIF 1.35 1.64 1.05 1.5 2.38 3.58 2.56 6.28 2.49 1.86 4.81 7.43 

## Analysis of Deviance Table
## 
## Model 1: Mcal ~ depth + om + gravel + sand + clay + arsenic + cadmium + 
##     chromium + copper + iron + manganese + mercury + lead + zinc
## Model 2: Mcal ~ depth + om + gravel + sand + arsenic + cadmium + chromium + 
##     copper + manganese + mercury + lead + zinc
##   Resid. Df Resid. Dev Df Deviance
## 1        96     613.58            
## 2        98     613.76 -2  -0.1812
Abundance of E. parma

This species is not characteristic of BSI, so low abundances can be a problem.

## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.45
Fitting generalized (poisson/log) linear model: Epar ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 5.242 1.546 3.392 0.000694 * * *
depth 0.002041 0.009914 0.2058 0.8369
om -2.607 0.5098 -5.114 3.148e-07 * * *
gravel -4.339 2.102 -2.064 0.03901 *
sand 2.573 1.12 2.297 0.02159 *
clay 2.36 5.038 0.4685 0.6394
arsenic 0.2437 0.2089 1.167 0.2433
cadmium -23.13 14.96 -1.546 0.1222
chromium 0.0586 0.03287 1.783 0.0746
copper 0.4406 0.136 3.239 0.001199 * *
iron 3.668e-06 1.017e-05 0.3606 0.7184
manganese -0.003072 0.001355 -2.267 0.02336 *
mercury 51.59 15.64 3.299 0.0009687 * * *
lead 0.07575 0.3469 0.2184 0.8271
zinc -0.1686 0.0623 -2.706 0.006817 * *
Variance Inflation Factors
  depth om gravel sand clay arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 1.37 1.44 1.22 2.28 1.14 1.89 3.35 4.48 4.91 1.63 3.44 2.88 4.96 7.46 

## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.44
Fitting generalized (poisson/log) linear model: Epar ~ om + gravel + sand + cadmium + chromium + copper + manganese + mercury + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 5.51 1.404 3.923 8.731e-05 * * *
om -2.662 0.4405 -6.042 1.52e-09 * * *
gravel -4.147 1.979 -2.096 0.03613 *
sand 2.356 0.999 2.358 0.01838 *
cadmium -22.1 12.03 -1.836 0.06631
chromium 0.03813 0.02042 1.867 0.06185
copper 0.4162 0.1259 3.306 0.000945 * * *
manganese -0.001972 0.0009234 -2.135 0.03274 *
mercury 52.51 12.94 4.057 4.97e-05 * * *
zinc -0.1436 0.05292 -2.713 0.006658 * *
Variance Inflation Factors
  om gravel sand cadmium chromium copper manganese mercury zinc
VIF 1.24 1.14 2.01 2.63 2.76 4.53 2.36 2.43 6.21 

## Analysis of Deviance Table
## 
## Model 1: Epar ~ depth + om + gravel + sand + clay + arsenic + cadmium + 
##     chromium + copper + iron + manganese + mercury + lead + zinc
## Model 2: Epar ~ om + gravel + sand + cadmium + chromium + copper + manganese + 
##     mercury + zinc
##   Resid. Df Resid. Dev Df Deviance
## 1        96     152.92            
## 2       101     155.25 -5  -2.3384
Abundance of Nematoda

This species is not characteristic of BSI, so low abundances can be a problem.

## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.63
Fitting generalized (poisson/log) linear model: Nema ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.319 0.478 0.6674 0.5045
depth 0.01861 0.003137 5.933 2.983e-09 * * *
om -2.169 0.148 -14.65 1.315e-48 * * *
gravel 4.004 0.3333 12.01 3.028e-33 * * *
sand 3.295 0.3576 9.214 3.136e-20 * * *
clay 1.027 1.555 0.6605 0.5089
arsenic 0.985 0.05475 17.99 2.234e-72 * * *
cadmium -20.23 5.646 -3.583 0.0003392 * * *
chromium 0.01349 0.00933 1.446 0.1481
copper -0.1429 0.05585 -2.559 0.01049 *
iron -4.467e-05 7.66e-06 -5.831 5.508e-09 * * *
manganese -0.0004546 0.000324 -1.403 0.1605
mercury 32.59 6.515 5.003 5.646e-07 * * *
lead -1.644 0.1239 -13.26 3.698e-40 * * *
zinc 0.1958 0.02238 8.748 2.175e-18 * * *
Variance Inflation Factors
  depth om gravel sand clay arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 2.02 1.63 2.32 2.3 1.09 3.64 4.87 4.8 8.77 2.86 3.96 2.35 7.31 11.5 

## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.63
Fitting generalized (poisson/log) linear model: Nema ~ depth + om + gravel + sand + arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.3942 0.4644 0.8488 0.396
depth 0.01855 0.003132 5.923 3.171e-09 * * *
om -2.172 0.1483 -14.65 1.352e-48 * * *
gravel 3.953 0.3246 12.18 4.095e-34 * * *
sand 3.219 0.3388 9.499 2.123e-21 * * *
arsenic 0.9787 0.05382 18.19 6.757e-74 * * *
cadmium -20.51 5.626 -3.646 0.000266 * * *
chromium 0.01395 0.009288 1.502 0.1332
copper -0.1449 0.05572 -2.6 0.009315 * *
iron -4.47e-05 7.65e-06 -5.843 5.141e-09 * * *
manganese -0.0004526 0.0003238 -1.398 0.1622
mercury 31.78 6.386 4.977 6.457e-07 * * *
lead -1.632 0.1222 -13.35 1.164e-40 * * *
zinc 0.1958 0.02239 8.746 2.204e-18 * * *
Variance Inflation Factors
  depth om gravel sand arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 2.02 1.63 2.26 2.17 3.58 4.85 4.78 8.75 2.85 3.96 2.3 7.22 11.5 

## Analysis of Deviance Table
## 
## Model 1: Nema ~ depth + om + gravel + sand + clay + arsenic + cadmium + 
##     chromium + copper + iron + manganese + mercury + lead + zinc
## Model 2: Nema ~ depth + om + gravel + sand + arsenic + cadmium + chromium + 
##     copper + iron + manganese + mercury + lead + zinc
##   Resid. Df Resid. Dev Df Deviance
## 1        96     1438.9            
## 2        97     1439.3 -1 -0.37093
Species richness
## FULL MODEL
## Adjusted R2 is: 0.19
Fitting linear model: S ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 20.89 4.908 4.255 4.853e-05 * * *
depth 0.05401 0.03024 1.786 0.07725
om 1.049 0.8738 1.2 0.233
gravel -0.202 4.225 -0.04782 0.962
sand 1.02 3.733 0.2732 0.7853
clay -0.4441 3.874 -0.1147 0.909
arsenic 0.1496 0.6386 0.2342 0.8153
cadmium -100.4 44.85 -2.239 0.02745 *
chromium 0.0484 0.0826 0.5859 0.5593
copper -0.2912 0.5598 -0.5202 0.6041
iron -9.849e-05 4.158e-05 -2.369 0.01985 *
manganese -0.002435 0.002383 -1.022 0.3094
mercury -26.03 58.91 -0.4419 0.6596
lead 1.508 1.158 1.302 0.1959
zinc 0.04078 0.21 0.1941 0.8465
Variance Inflation Factors
  depth om gravel sand clay arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 1.33 1.98 1.07 2.03 1.15 2.22 2.8 3.15 6.34 1.45 2.91 1.8 4.75 7.49 

## REDUCED MODEL
## Adjusted R2 is: 0.25
Fitting linear model: S ~ depth + cadmium + iron + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 20.95 3 6.984 2.616e-10 * * *
depth 0.07349 0.02407 3.053 0.002862 * *
cadmium -67.45 24.81 -2.718 0.007665 * *
iron -0.0001053 3.024e-05 -3.481 0.0007266 * * *
lead 0.8049 0.386 2.085 0.03946 *
Variance Inflation Factors
  depth cadmium iron lead
VIF 1.1 1.6 1.09 1.64 

## Analysis of Variance Table
## 
## Model 1: S ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + 
##     copper + iron + manganese + mercury + lead + zinc
## Model 2: S ~ depth + cadmium + iron + lead
##   Res.Df    RSS  Df Sum of Sq      F Pr(>F)
## 1     96 2100.5                            
## 2    106 2168.3 -10   -67.791 0.3098  0.977

## RMSE for the full model: 5.052289
## RMSE for the reduced model: 4.642693
Total abundance
## FULL MODEL
## Adjusted R2 is: 0.05
Fitting linear model: N ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 163.5 81.07 2.017 0.04656 *
depth -0.8406 0.4889 -1.72 0.08876
om 3.534 14.16 0.2496 0.8034
gravel 39.94 68.72 0.5812 0.5625
sand 20.62 61.43 0.3357 0.7379
clay -74.72 62.74 -1.191 0.2367
arsenic 4.502 10.34 0.4353 0.6643
cadmium -1300 728.6 -1.784 0.07762
chromium 0.4344 1.335 0.3254 0.7456
copper -8.75 9.16 -0.9552 0.3419
iron -0.0005995 0.0006716 -0.8926 0.3743
manganese -0.04376 0.03849 -1.137 0.2584
mercury -899.4 951.6 -0.9452 0.3469
lead 38.53 18.72 2.058 0.04232 *
zinc 1.324 3.453 0.3833 0.7023
Variance Inflation Factors
  depth om gravel sand clay arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 1.32 1.98 1.08 2.05 1.16 2.22 2.8 3.14 6.4 1.45 2.91 1.8 4.73 7.56 

## REDUCED MODEL
## Adjusted R2 is: 0.09
Fitting linear model: N ~ depth + cadmium + manganese + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 170 43.85 3.876 0.000185 * * *
depth -0.7761 0.3971 -1.955 0.0533
cadmium -773.9 406.7 -1.903 0.05977
manganese -0.05945 0.01982 -3 0.003375 * *
lead 23.04 7.627 3.021 0.003169 * *
Variance Inflation Factors
  depth cadmium manganese lead
VIF 1.1 1.59 1.52 1.96 

## Analysis of Variance Table
## 
## Model 1: N ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + 
##     copper + iron + manganese + mercury + lead + zinc
## Model 2: N ~ depth + cadmium + manganese + lead
##   Res.Df    RSS  Df Sum of Sq      F Pr(>F)
## 1     95 542243                            
## 2    105 578171 -10    -35928 0.6295 0.7853

## RMSE for the full model: 83.17193
## RMSE for the reduced model: 75.35638
Shannon index
## FULL MODEL
## Adjusted R2 is: 0.36
Fitting linear model: H ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.186 0.4506 4.852 4.73e-06 * * *
depth 0.01003 0.002776 3.614 0.0004829 * * *
om 0.1624 0.08021 2.024 0.04574 *
gravel -0.2883 0.3878 -0.7434 0.4591
sand 0.1327 0.3427 0.3873 0.6994
clay 0.6397 0.3556 1.799 0.07518
arsenic -0.06008 0.05862 -1.025 0.308
cadmium -4.535 4.117 -1.101 0.2734
chromium -0.003811 0.007582 -0.5026 0.6164
copper -0.01353 0.05139 -0.2634 0.7928
iron -6.19e-06 3.817e-06 -1.622 0.1081
manganese 7.63e-05 0.0002187 0.3488 0.728
mercury 0.225 5.408 0.04162 0.9669
lead 0.0373 0.1063 0.3509 0.7265
zinc 0.003476 0.01928 0.1803 0.8573
Variance Inflation Factors
  depth om gravel sand clay arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 1.33 1.98 1.07 2.03 1.15 2.22 2.8 3.15 6.34 1.45 2.91 1.8 4.75 7.49 

## REDUCED MODEL
## Adjusted R2 is: 0.4
Fitting linear model: H ~ depth + om + clay + cadmium + iron
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.398 0.2877 8.335 3.204e-13 * * *
depth 0.009826 0.002459 3.996 0.0001201 * * *
om 0.1437 0.04872 2.949 0.003929 * *
clay 0.4838 0.3047 1.588 0.1153
cadmium -5.308 1.828 -2.904 0.004493 * *
iron -8.055e-06 2.755e-06 -2.924 0.004237 * *
Variance Inflation Factors
  depth om clay cadmium iron
VIF 1.22 1.24 1.02 1.28 1.08 

## Analysis of Variance Table
## 
## Model 1: H ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + 
##     copper + iron + manganese + mercury + lead + zinc
## Model 2: H ~ depth + om + clay + cadmium + iron
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1     96 17.700                           
## 2    105 18.252 -9  -0.55204 0.3327 0.9621

## RMSE for the full model: 0.514536
## RMSE for the reduced model: 0.4369674
Piélou’s evenness
## FULL MODEL
## Adjusted R2 is: 0.28
Fitting linear model: J ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.7022 0.1567 4.48 2.061e-05 * * *
depth 0.002801 0.0009657 2.9 0.004621 * *
om 0.00709 0.0279 0.2541 0.8
gravel -0.1254 0.1349 -0.9297 0.3548
sand -0.09292 0.1192 -0.7794 0.4376
clay 0.2574 0.1237 2.081 0.04009 *
arsenic -0.04734 0.02039 -2.321 0.02239 *
cadmium 2.07 1.432 1.445 0.1517
chromium -0.003906 0.002638 -1.481 0.1419
copper 0.0001094 0.01788 0.006118 0.9951
iron 1.052e-06 1.328e-06 0.7924 0.4301
manganese 0.0001575 7.61e-05 2.07 0.04111 *
mercury 0.3132 1.881 0.1665 0.8681
lead -0.03207 0.03698 -0.8674 0.3879
zinc 5.44e-05 0.006707 0.008111 0.9935
Variance Inflation Factors
  depth om gravel sand clay arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 1.33 1.98 1.07 2.03 1.15 2.22 2.8 3.15 6.34 1.45 2.91 1.8 4.75 7.49 

## REDUCED MODEL
## Adjusted R2 is: 0.3
Fitting linear model: J ~ depth + clay + arsenic + cadmium + chromium + manganese + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.6259 0.0889 7.04 2.211e-10 * * *
depth 0.003258 0.0008167 3.989 0.0001244 * * *
clay 0.302 0.1102 2.741 0.007218 * *
arsenic -0.04566 0.01791 -2.55 0.01224 *
cadmium 2.214 0.8719 2.539 0.01261 *
chromium -0.003954 0.00189 -2.092 0.03889 *
manganese 0.0001903 4.996e-05 3.809 0.0002378 * * *
lead -0.02602 0.01894 -1.374 0.1726
Variance Inflation Factors
  depth clay arsenic cadmium chromium manganese lead
VIF 1.15 1.05 1.98 1.74 2.3 1.94 2.48 

## Analysis of Variance Table
## 
## Model 1: J ~ depth + om + gravel + sand + clay + arsenic + cadmium + chromium + 
##     copper + iron + manganese + mercury + lead + zinc
## Model 2: J ~ depth + clay + arsenic + cadmium + chromium + manganese + 
##     lead
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1     96 2.1419                           
## 2    103 2.2114 -7 -0.069546 0.4453 0.8711

## RMSE for the full model: 0.175133
## RMSE for the reduced model: 0.1485803

Parameters

i) Simple regressions

These analyses have been do to explore the relationships between variables. As it is a huge number of results to interpret, only multiple regressions will be included in the article (see below).

Adjusted R-squared of simple regressions with parameters at BSI and MR
  depth om gravel sand silt clay
Bneo 0.02066 0.006557 0.000244 0.01557 0.05216 -0.001626
Mcal -0.006558 0.003274 0.0002456 0.0004419 -0.006683 -0.007314
Epar 0.03402 0.1292 -0.006223 0.179 0.1513 -0.005043
Nema -0.007161 0.08987 0.05438 0.03268 0.07902 -0.003509
S 0.1555 -0.004351 -0.002872 0.006156 0.003318 -0.006916
N 0.02062 0.007067 -0.0005827 -0.002467 -0.004443 0.004763
H 0.2897 0.03561 -0.008062 0.0246 0.007538 0.006524
J 0.1198 0.02409 -0.005515 0.006385 -0.00433 0.02011
p-values of simple regressions with parameters at BSI and MR
  depth om gravel sand silt clay
Bneo 0.05883 0.1792 0.312 0.08696 0.00581 0.3737
Mcal 0.6673 0.2372 0.312 0.3063 0.6807 0.7617
Epar 0.02174 2.113e-05 0.6347 4.78e-07 4.035e-06 0.5426
Nema 0.7393 0.0003812 0.004948 0.02398 0.0008364 0.4545
S 2.915e-06 0.4996 0.4245 0.1853 0.2363 0.7074
N 0.05901 0.1717 0.3375 0.407 0.505 0.2085
H 4.833e-11 0.01935 0.9874 0.04369 0.1651 0.1797
J 4.26e-05 0.04539 0.576 0.1818 0.4984 0.06133
ii) Multiple regressions

To account for the high number of zeros in the characteristic species abundances, we used both Poisson and Zero-Inflated Poisson models for these regressions.

This table is summarizing the significative relationships with abiotic parameters as predictors, obtained with the AIC-reduced models from the multiple regressions below (ZIP not included)

Predictor Bneo Mcal Epar Nema S N H J
depth - + + + +
om - - - - -
gravel - - - +
sand -
silt - + +
clay - - + +
McFadden Pseudo-R2 0.21 0.04 0.42 0.45
Adjusted-R2 0.16 0.01 0.30 0.14
Abundance of B. neotena
## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.21
Fitting generalized (poisson/log) linear model: Bneo ~ depth + om + gravel + sand + clay
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 4.984 0.1049 47.49 0 * * *
depth -0.02219 0.00147 -15.1 1.696e-51 * * *
om -0.1914 0.02967 -6.45 1.121e-10 * * *
gravel -8.638 1.04 -8.304 1.006e-16 * * *
sand -2.384 0.1329 -17.95 5.224e-72 * * *
clay -6.552 0.5505 -11.9 1.141e-32 * * *
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.03 1.42 1.02 1.42 1.07 
## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.21
Fitting generalized (poisson/log) linear model: Bneo ~ depth + om + gravel + sand + clay
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 4.984 0.1049 47.49 0 * * *
depth -0.02219 0.00147 -15.1 1.696e-51 * * *
om -0.1914 0.02967 -6.45 1.121e-10 * * *
gravel -8.638 1.04 -8.304 1.006e-16 * * *
sand -2.384 0.1329 -17.95 5.224e-72 * * *
clay -6.552 0.5505 -11.9 1.141e-32 * * *
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.03 1.42 1.02 1.42 1.07 

## Analysis of Deviance Table
## 
## Model 1: Bneo ~ depth + om + gravel + sand + clay
## Model 2: Bneo ~ depth + om + gravel + sand + clay
##   Resid. Df Resid. Dev Df Deviance
## 1       119     3227.1            
## 2       119     3227.1  0        0
## FULL MODEL (Zero-Inflated Poisson)
Fitting corresponding ZIP generalized linear model (counts)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 3.512 0.07284 48.21 0
depth -0.03064 0.001526 -20.08 1.133e-89
om -0.5021 0.04009 -12.53 5.426e-36
gravel -3.35 0.9649 -3.472 0.0005171
silt 2.555 0.1869 13.67 1.571e-42
clay -3.561 0.531 -6.705 2.012e-11
Fitting corresponding ZIP generalized linear model (zeros)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 1.89 0.5951 3.176 0.001491
depth -0.0166 0.01513 -1.097 0.2726
om -1.855 0.6182 -3.001 0.002688
gravel -0.1986 2.515 -0.07897 0.9371
silt -1.578 1.357 -1.163 0.245
clay 2.841 1.983 1.433 0.152 
## REDUCED MODEL
Fitting corresponding ZIP generalized linear model (counts)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 3.512 0.07284 48.21 0
depth -0.03064 0.001526 -20.08 1.133e-89
om -0.5021 0.04009 -12.53 5.426e-36
gravel -3.35 0.9649 -3.472 0.0005171
silt 2.555 0.1869 13.67 1.571e-42
clay -3.561 0.531 -6.705 2.012e-11
Fitting corresponding ZIP generalized linear model (zeros)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 1.89 0.5951 3.176 0.001491
depth -0.0166 0.01513 -1.097 0.2726
om -1.855 0.6182 -3.001 0.002688
gravel -0.1986 2.515 -0.07897 0.9371
silt -1.578 1.357 -1.163 0.245
clay 2.841 1.983 1.433 0.152
Abundance of M. calcarea
## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.04
Fitting generalized (poisson/log) linear model: Mcal ~ depth + om + gravel + sand + clay
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 2.1 0.2266 9.268 1.89e-20 * * *
depth -0.0002055 0.002301 -0.08931 0.9288
om -0.2879 0.0709 -4.06 4.907e-05 * * *
gravel -2.973 0.6931 -4.289 1.792e-05 * * *
sand -0.2227 0.2539 -0.8771 0.3804
clay -0.4611 0.4535 -1.017 0.3093
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.05 1.49 1.03 1.55 1.14 

## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.04
Fitting generalized (poisson/log) linear model: Mcal ~ om + gravel
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 1.909 0.07337 26.01 3.499e-149 * * *
om -0.2511 0.04814 -5.216 1.833e-07 * * *
gravel -2.872 0.6858 -4.188 2.82e-05 * * *
Variance Inflation Factors
  om gravel
VIF 1.01 1.01 

## Analysis of Deviance Table
## 
## Model 1: Mcal ~ depth + om + gravel + sand + clay
## Model 2: Mcal ~ om + gravel
##   Resid. Df Resid. Dev Df Deviance
## 1       120     1078.1            
## 2       123     1079.4 -3  -1.3235
## FULL MODEL (Zero-Inflated Poisson)
Fitting corresponding ZIP generalized linear model (counts)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 3 0.1167 25.7 1.091e-145
depth -0.007223 0.002472 -2.922 0.003483
om -0.3606 0.08189 -4.403 1.067e-05
gravel -2.99 0.8165 -3.662 0.0002505
silt -0.9106 0.3105 -2.932 0.003364
clay 0.7077 0.4151 1.705 0.08823
Fitting corresponding ZIP generalized linear model (zeros)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.9476 0.5583 1.697 0.08966
depth -0.01665 0.01507 -1.105 0.2692
om -0.8454 0.6339 -1.334 0.1823
gravel 0.5623 2.24 0.251 0.8018
silt -1.634 1.33 -1.229 0.2192
clay 3.328 1.913 1.739 0.08201 
## REDUCED MODEL
Fitting corresponding ZIP generalized linear model (counts)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 3 0.1167 25.7 1.091e-145
depth -0.007223 0.002472 -2.922 0.003483
om -0.3606 0.08189 -4.403 1.067e-05
gravel -2.99 0.8165 -3.662 0.0002505
silt -0.9106 0.3105 -2.932 0.003364
clay 0.7077 0.4151 1.705 0.08823
Fitting corresponding ZIP generalized linear model (zeros)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.9476 0.5583 1.697 0.08966
depth -0.01665 0.01507 -1.105 0.2692
om -0.8454 0.6339 -1.334 0.1823
gravel 0.5623 2.24 0.251 0.8018
silt -1.634 1.33 -1.229 0.2192
clay 3.328 1.913 1.739 0.08201
Abundance of E. parma

Abundances are very low for this species.

## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.42
Fitting generalized (poisson/log) linear model: Epar ~ depth + om + gravel + silt + clay
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 2.765 0.1565 17.67 7.629e-70 * * *
depth 0.002571 0.005288 0.4862 0.6268
om -2.29 0.334 -6.855 7.133e-12 * * *
gravel -5.524 1.272 -4.341 1.419e-05 * * *
silt -2.361 0.5532 -4.268 1.969e-05 * * *
clay 2.038 1.26 1.618 0.1057
Variance Inflation Factors
  depth om gravel silt clay
VIF 1.15 1.35 1.03 1.22 1.1 

## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.42
Fitting generalized (poisson/log) linear model: Epar ~ om + gravel + silt
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 2.749 0.1526 18.02 1.426e-72 * * *
om -2.189 0.3093 -7.079 1.45e-12 * * *
gravel -5.412 1.261 -4.29 1.784e-05 * * *
silt -2.195 0.5615 -3.909 9.283e-05 * * *
Variance Inflation Factors
  om gravel silt
VIF 1.23 1.01 1.23 

## Analysis of Deviance Table
## 
## Model 1: Epar ~ depth + om + gravel + silt + clay
## Model 2: Epar ~ om + gravel + silt
##   Resid. Df Resid. Dev Df Deviance
## 1       120     314.51            
## 2       122     316.61 -2  -2.1088
## FULL MODEL (Zero-Inflated Poisson)
Fitting corresponding ZIP generalized linear model (counts)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 2.891 0.2016 14.34 1.268e-46
depth -0.01208 0.00796 -1.517 0.1292
om -1.264 0.3898 -3.242 0.001188
gravel -2.964 1.284 -2.309 0.02093
silt -0.7471 0.4346 -1.719 0.08561
clay -0.5372 1.672 -0.3212 0.748
Fitting corresponding ZIP generalized linear model (zeros)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) -1.162 0.5966 -1.948 0.05141
depth -0.002824 0.01584 -0.1783 0.8585
om 1.522 0.9029 1.685 0.09194
gravel 1.491 2.53 0.5894 0.5556
silt 2.008 1.996 1.006 0.3143
clay -5.186 4.818 -1.076 0.2817 
## REDUCED MODEL
Fitting corresponding ZIP generalized linear model (counts)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 2.764 0.1762 15.69 1.863e-55
om -1.428 0.3651 -3.912 9.149e-05
gravel -3.484 1.357 -2.567 0.01027
silt -0.875 0.4536 -1.929 0.05375
Fitting corresponding ZIP generalized linear model (zeros)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) -1.097 0.546 -2.01 0.04446
om 1.537 0.7136 2.154 0.03122
gravel 1.3 2.473 0.5256 0.5992
silt 1.165 1.505 0.7741 0.4389
Abundance of Nematoda

Abundances are very low for this species.

## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.45
Fitting generalized (poisson/log) linear model: Nema ~ depth + om + gravel + silt + clay
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 3.811 0.06576 57.95 0 * * *
depth 0.0209 0.001782 11.73 8.912e-32 * * *
om -2.976 0.1209 -24.63 6.653e-134 * * *
gravel 1.322 0.1542 8.578 9.628e-18 * * *
silt 0.8836 0.1787 4.945 7.634e-07 * * *
clay -5.458 1.417 -3.852 0.0001173 * * *
Variance Inflation Factors
  depth om gravel silt clay
VIF 1.18 1.34 1.18 1.28 1.02 
## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.45
Fitting generalized (poisson/log) linear model: Nema ~ depth + om + gravel + silt + clay
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 3.811 0.06576 57.95 0 * * *
depth 0.0209 0.001782 11.73 8.912e-32 * * *
om -2.976 0.1209 -24.63 6.653e-134 * * *
gravel 1.322 0.1542 8.578 9.628e-18 * * *
silt 0.8836 0.1787 4.945 7.634e-07 * * *
clay -5.458 1.417 -3.852 0.0001173 * * *
Variance Inflation Factors
  depth om gravel silt clay
VIF 1.18 1.34 1.18 1.28 1.02 

## Analysis of Deviance Table
## 
## Model 1: Nema ~ depth + om + gravel + silt + clay
## Model 2: Nema ~ depth + om + gravel + silt + clay
##   Resid. Df Resid. Dev Df Deviance
## 1       120     2464.2            
## 2       120     2464.2  0        0
## FULL MODEL (Zero-Inflated Poisson)
Fitting corresponding ZIP generalized linear model (counts)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 3.957 0.06734 58.76 0
depth 0.01441 0.001915 7.525 5.263e-14
om -2.49 0.1439 -17.31 4.257e-67
gravel 1.098 0.1574 6.978 2.986e-12
silt 1.034 0.1626 6.363 1.974e-10
clay -3.788 1.47 -2.576 0.01
Fitting corresponding ZIP generalized linear model (zeros)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) -1.343 0.5456 -2.461 0.01387
depth -0.05033 0.01583 -3.18 0.001472
om 2.196 0.6936 3.167 0.001542
gravel -0.6259 2.952 -0.212 0.8321
silt 0.2631 1.622 0.1622 0.8711
clay -0.9465 3.245 -0.2916 0.7706 
## REDUCED MODEL
Fitting corresponding ZIP generalized linear model (counts)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 3.957 0.06734 58.76 0
depth 0.01441 0.001915 7.525 5.263e-14
om -2.49 0.1439 -17.31 4.257e-67
gravel 1.098 0.1574 6.978 2.986e-12
silt 1.034 0.1626 6.363 1.974e-10
clay -3.788 1.47 -2.576 0.01
Fitting corresponding ZIP generalized linear model (zeros)
  Estimate Std. Error z value Pr(>|z|)
(Intercept) -1.343 0.5456 -2.461 0.01387
depth -0.05033 0.01583 -3.18 0.001472
om 2.196 0.6936 3.167 0.001542
gravel -0.6259 2.952 -0.212 0.8321
silt 0.2631 1.622 0.1622 0.8711
clay -0.9465 3.245 -0.2916 0.7706
Species richness
## FULL MODEL
## Adjusted R2 is: 0.15
Fitting linear model: S ~ depth + om + gravel + silt + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 8.419 1.032 8.157 3.841e-13 * * *
depth 0.1088 0.02274 4.787 4.873e-06 * * *
om -0.8085 0.6478 -1.248 0.2144
gravel 4.139 3.77 1.098 0.2745
silt 4.209 2.67 1.576 0.1175
clay 1.003 3.42 0.2932 0.7699
Variance Inflation Factors
  depth om gravel silt clay
VIF 1.02 1.54 1.09 1.59 1.07 

## REDUCED MODEL
## Adjusted R2 is: 0.16
Fitting linear model: S ~ depth
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 9.166 0.756 12.12 8.564e-23 * * *
depth 0.1082 0.02208 4.901 2.915e-06 * * *
Variance Inflation Factors
  depth
VIF 1 

## Analysis of Variance Table
## 
## Model 1: S ~ depth + om + gravel + silt + clay
## Model 2: S ~ depth
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1    120 2696.2                           
## 2    124 2760.6 -4   -64.401 0.7166 0.5822

## RMSE for the full model: 4.85609
## RMSE for the reduced model: 4.74687
Total abundance
## FULL MODEL
## Adjusted R2 is: 0.02
Fitting linear model: N ~ depth + om + gravel + silt + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 94.39 16.42 5.747 7.119e-08 * * *
depth -0.4541 0.3587 -1.266 0.208
om -13.69 10.16 -1.348 0.1801
gravel 75.9 59.1 1.284 0.2015
silt 60.06 41.98 1.431 0.1552
clay -50.95 53.63 -0.95 0.344
Variance Inflation Factors
  depth om gravel silt clay
VIF 1.02 1.54 1.09 1.59 1.07 

## REDUCED MODEL
## Adjusted R2 is: 0.01
Fitting linear model: N ~ om + silt
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 88.22 12.85 6.863 3.004e-10 * * *
om -15.98 9.33 -1.713 0.08923
silt 54.03 37.31 1.448 0.1501
Variance Inflation Factors
  om silt
VIF 1.4 1.4 

## Analysis of Variance Table
## 
## Model 1: N ~ depth + om + gravel + silt + clay
## Model 2: N ~ om + silt
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1    119 656840                           
## 2    122 680015 -3    -23174 1.3995 0.2464

## RMSE for the full model: 76.2312
## RMSE for the reduced model: 75.91861
Shannon index
## FULL MODEL
## Adjusted R2 is: 0.29
Fitting linear model: H ~ depth + om + gravel + silt + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.26 0.09796 12.87 2.442e-24 * * *
depth 0.01491 0.002158 6.911 2.478e-10 * * *
om 0.03575 0.06148 0.5815 0.562
gravel -0.06505 0.3578 -0.1818 0.856
silt 0.05329 0.2534 0.2103 0.8338
clay 0.5161 0.3245 1.59 0.1144
Variance Inflation Factors
  depth om gravel silt clay
VIF 1.02 1.54 1.09 1.59 1.07 

## REDUCED MODEL
## Adjusted R2 is: 0.3
Fitting linear model: H ~ depth + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.316 0.0723 18.2 1.049e-36 * * *
depth 0.01534 0.002093 7.328 2.695e-11 * * *
clay 0.5516 0.302 1.827 0.07019
Variance Inflation Factors
  depth clay
VIF 1 1 

## Analysis of Variance Table
## 
## Model 1: H ~ depth + om + gravel + silt + clay
## Model 2: H ~ depth + clay
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1    120 24.285                           
## 2    123 24.568 -3  -0.28358 0.4671 0.7058

## RMSE for the full model: 0.4735333
## RMSE for the reduced model: 0.4516556
Piélou’s evenness
## FULL MODEL
## Adjusted R2 is: 0.14
Fitting linear model: J ~ depth + om + gravel + silt + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.6347 0.03476 18.26 1.914e-36 * * *
depth 0.003122 0.0007656 4.078 8.212e-05 * * *
om 0.02604 0.02181 1.194 0.2349
gravel -0.1161 0.1269 -0.9142 0.3624
silt -0.06689 0.08991 -0.744 0.4583
clay 0.1838 0.1151 1.596 0.113
Variance Inflation Factors
  depth om gravel silt clay
VIF 1.02 1.54 1.09 1.59 1.07 

## REDUCED MODEL
## Adjusted R2 is: 0.14
Fitting linear model: J ~ depth + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.6352 0.02572 24.69 1.692e-49 * * *
depth 0.003252 0.0007445 4.367 2.641e-05 * * *
clay 0.2309 0.1074 2.149 0.03357 *
Variance Inflation Factors
  depth clay
VIF 1 1 

## Analysis of Variance Table
## 
## Model 1: J ~ depth + om + gravel + silt + clay
## Model 2: J ~ depth + clay
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1    120 3.0571                           
## 2    123 3.1099 -3 -0.052755 0.6903 0.5597

## RMSE for the full model: 0.1676339
## RMSE for the reduced model: 0.1598252

Metals

i) Simple regressions

These analyses have been do to explore the relationships between variables. As it is a huge number of results to interpret, only multiple regressions will be included in the article (see below).

Adjusted R-squared of simple regressions with metals at BSI
  arsenic cadmium chromium copper iron manganese mercury lead zinc
Bneo 0.02658 0.05792 0.03385 0.05284 -0.007023 -0.007612 0.01923 0.06944 0.09318
Mcal 0.04531 0.06456 0.03261 0.004907 -0.007508 0.03006 0.06613 0.06284 0.01137
Epar 0.01663 0.03255 0.0243 0.02964 0.002643 0.02949 -0.005137 0.02621 0.03627
Nema -0.009159 0.002577 -0.005425 0.005019 -0.003006 0.003156 0.0006913 0.00456 0.001428
S 0.01662 0.1027 0.118 0.05078 0.1085 0.07493 0.002283 0.03525 0.06309
N -0.006692 -0.004822 -0.008279 -0.003719 -0.001477 -0.002075 -0.008671 0.005467 0.005936
H 0.05214 0.1186 0.1572 0.09365 0.06617 0.0456 0.00858 0.086 0.1237
J 0.06716 0.02249 0.06068 0.07452 -0.00309 -0.0007157 0.004781 0.07601 0.08759
p-values of simple regressions with metals at BSI
  arsenic cadmium chromium copper iron manganese mercury lead zinc
Bneo 0.04787 0.006292 0.02968 0.008709 0.6304 0.6818 0.0784 0.003012 0.000658
Mcal 0.01413 0.004114 0.03219 0.2169 0.6719 0.03805 0.003723 0.004594 0.1352
Epar 0.09367 0.03233 0.05573 0.03912 0.2583 0.03952 0.5096 0.04909 0.02535
Nema 0.9675 0.2596 0.5251 0.2151 0.4147 0.2481 0.3019 0.2227 0.2844
S 0.09373 0.0003555 0.0001322 0.009938 0.000244 0.002121 0.2657 0.0271 0.004519
N 0.6052 0.4935 0.7563 0.4431 0.3621 0.3815 0.8161 0.2079 0.2008
H 0.009112 0.0001268 9.806e-06 0.0006383 0.003713 0.01387 0.1652 0.001044 9.084e-05
J 0.003486 0.0629 0.005274 0.002177 0.4179 0.3393 0.219 0.001979 0.0009426
ii) Multiple regressions

ZIP models are “computationally” singular, so they have not been computed here.

This table is summarizing the significative relationships with abiotic parameters as predictors, obtained with the AIC-reduced models from the multiple regressions below (ZIP not included)

Predictor Bneo Mcal Epar Nema S N H J
arsenic + - +
cadmium - - - - - +
chromium - + + - -
copper - - + - -
iron - - + - -
manganese - + - - - + +
mercury + - +
lead - + - + + -
zinc + - -
McFadden Pseudo-R2 0.29 0.35 0.21 0.27
Adjusted-R2 0.19 0.09 0.18 0.16
Abundance of B. neotena
## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.29
Fitting generalized (poisson/log) linear model: Bneo ~ arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 1.229 0.1704 7.209 5.618e-13 * * *
arsenic 0.04042 0.02611 1.548 0.1215
cadmium -3.637 2.408 -1.51 0.1309
chromium -0.004222 0.005658 -0.7462 0.4555
copper -0.05152 0.02482 -2.075 0.03795 *
iron -1.498e-05 5.514e-06 -2.718 0.006574 * *
manganese -0.001245 0.0001363 -9.134 6.574e-20 * * *
mercury 46.95 3.501 13.41 5.253e-41 * * *
lead 0.0005271 0.05816 0.009063 0.9928
zinc 0.06668 0.009634 6.921 4.484e-12 * * *
Variance Inflation Factors
  arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 2.09 3.34 4.45 5.87 2.93 3.12 1.83 5.01 8.13 

## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.29
Fitting generalized (poisson/log) linear model: Bneo ~ arsenic + cadmium + copper + iron + manganese + mercury + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 1.243 0.1685 7.378 1.603e-13 * * *
arsenic 0.04415 0.0225 1.962 0.04981 *
cadmium -4.035 2.255 -1.789 0.07361
copper -0.0537 0.02417 -2.221 0.02632 *
iron -1.559e-05 5.3e-06 -2.941 0.003271 * *
manganese -0.001281 0.0001261 -10.16 3.01e-24 * * *
mercury 46.59 3.163 14.73 4.105e-49 * * *
zinc 0.0649 0.008859 7.326 2.366e-13 * * *
Variance Inflation Factors
  arsenic cadmium copper iron manganese mercury zinc
VIF 1.8 3.12 5.72 2.81 2.87 1.65 7.48 

## Analysis of Deviance Table
## 
## Model 1: Bneo ~ arsenic + cadmium + chromium + copper + iron + manganese + 
##     mercury + lead + zinc
## Model 2: Bneo ~ arsenic + cadmium + copper + iron + manganese + mercury + 
##     zinc
##   Resid. Df Resid. Dev Df Deviance
## 1       100     2492.1            
## 2       102     2492.7 -2 -0.56019
Abundance of M. calcarea
## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.35
Fitting generalized (poisson/log) linear model: Mcal ~ arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 3.353 0.3032 11.06 1.944e-28 * * *
arsenic -0.2509 0.1085 -2.313 0.02075 *
cadmium -27.25 6.075 -4.485 7.294e-06 * * *
chromium -0.02745 0.007202 -3.812 0.0001378 * * *
copper -0.1298 0.06252 -2.076 0.03789 *
iron -6.484e-06 3.214e-06 -2.017 0.04365 *
manganese 0.001112 0.0002683 4.144 3.418e-05 * * *
mercury -14.97 8.212 -1.823 0.06831
lead -0.5097 0.141 -3.615 0.0003005 * * *
zinc 0.1243 0.02138 5.816 6.016e-09 * * *
Variance Inflation Factors
  arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 2.39 3.48 2.29 6.41 1.37 2.36 1.87 4.78 7.07 
## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.35
Fitting generalized (poisson/log) linear model: Mcal ~ arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 3.353 0.3032 11.06 1.944e-28 * * *
arsenic -0.2509 0.1085 -2.313 0.02075 *
cadmium -27.25 6.075 -4.485 7.294e-06 * * *
chromium -0.02745 0.007202 -3.812 0.0001378 * * *
copper -0.1298 0.06252 -2.076 0.03789 *
iron -6.484e-06 3.214e-06 -2.017 0.04365 *
manganese 0.001112 0.0002683 4.144 3.418e-05 * * *
mercury -14.97 8.212 -1.823 0.06831
lead -0.5097 0.141 -3.615 0.0003005 * * *
zinc 0.1243 0.02138 5.816 6.016e-09 * * *
Variance Inflation Factors
  arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 2.39 3.48 2.29 6.41 1.37 2.36 1.87 4.78 7.07 

## Analysis of Deviance Table
## 
## Model 1: Mcal ~ arsenic + cadmium + chromium + copper + iron + manganese + 
##     mercury + lead + zinc
## Model 2: Mcal ~ arsenic + cadmium + chromium + copper + iron + manganese + 
##     mercury + lead + zinc
##   Resid. Df Resid. Dev Df Deviance
## 1        99     426.71            
## 2        99     426.71  0        0
Abundance of E. parma

Abundances are very low for this species.

## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.21
Fitting generalized (poisson/log) linear model: Epar ~ arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 4.382 1.018 4.306 1.661e-05 * * *
arsenic 0.2146 0.2492 0.861 0.3892
cadmium -13.9 14.14 -0.9828 0.3257
chromium 0.0962 0.02354 4.087 4.376e-05 * * *
copper 0.3466 0.145 2.391 0.01681 *
iron 3.684e-05 1.088e-05 3.385 0.0007112 * * *
manganese -0.004212 0.001042 -4.042 5.31e-05 * * *
mercury 33.95 10.83 3.136 0.001714 * *
lead 0.5873 0.275 2.136 0.03268 *
zinc -0.2496 0.062 -4.027 5.657e-05 * * *
Variance Inflation Factors
  arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 1.89 2.95 3.34 5.36 1.57 3.01 1.6 3.65 7.05 

## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.21
Fitting generalized (poisson/log) linear model: Epar ~ chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 4.207 0.9844 4.274 1.923e-05 * * *
chromium 0.07753 0.0189 4.102 4.103e-05 * * *
copper 0.3992 0.1129 3.537 0.0004045 * * *
iron 3.684e-05 1.084e-05 3.399 0.0006771 * * *
manganese -0.003658 0.000816 -4.483 7.377e-06 * * *
mercury 34.39 10.72 3.209 0.001332 * *
lead 0.5215 0.2177 2.396 0.01659 *
zinc -0.2632 0.05351 -4.919 8.677e-07 * * *
Variance Inflation Factors
  chromium copper iron manganese mercury lead zinc
VIF 2.75 4.09 1.57 2.36 1.59 2.81 5.92 

## Analysis of Deviance Table
## 
## Model 1: Epar ~ arsenic + cadmium + chromium + copper + iron + manganese + 
##     mercury + lead + zinc
## Model 2: Epar ~ chromium + copper + iron + manganese + mercury + lead + 
##     zinc
##   Resid. Df Resid. Dev Df Deviance
## 1       100     243.07            
## 2       102     245.43 -2  -2.3615
Abundance of Nematoda

Abundances are very low for this species.

## FULL MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.27
Fitting generalized (poisson/log) linear model: Nema ~ arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 4.014 0.1948 20.61 2.133e-94 * * *
arsenic 0.8225 0.03743 21.97 5.033e-107 * * *
cadmium -20.45 3.977 -5.14 2.741e-07 * * *
chromium 0.1202 0.007634 15.75 7.36e-56 * * *
copper -0.05707 0.0249 -2.292 0.0219 *
iron -4.905e-05 7.282e-06 -6.735 1.64e-11 * * *
manganese -0.003388 0.0002613 -12.97 1.912e-38 * * *
mercury 0.7612 4.554 0.1671 0.8673
lead -0.4546 0.09573 -4.749 2.048e-06 * * *
Variance Inflation Factors
  arsenic cadmium chromium copper iron manganese mercury lead
VIF 2.59 3.76 4.33 4.14 3.11 3.34 1.49 5.96 

## REDUCED MODEL (Poisson)
## McFadden's Pseudo-R2 is: 0.27
Fitting generalized (poisson/log) linear model: Nema ~ arsenic + cadmium + chromium + copper + iron + manganese + lead
  Estimate Std. Error z value Pr(>|z|)
(Intercept) 4.019 0.1924 20.9 5.881e-97 * * *
arsenic 0.8217 0.03716 22.11 2.544e-108 * * *
cadmium -20.37 3.95 -5.157 2.504e-07 * * *
chromium 0.1201 0.007616 15.77 4.82e-56 * * *
copper -0.05756 0.02474 -2.327 0.01997 *
iron -4.928e-05 7.137e-06 -6.905 5.005e-12 * * *
manganese -0.003387 0.0002611 -12.97 1.836e-38 * * *
lead -0.4499 0.09154 -4.915 8.888e-07 * * *
Variance Inflation Factors
  arsenic cadmium chromium copper iron manganese lead
VIF 2.57 3.74 4.32 4.11 3.05 3.33 5.7 

## Analysis of Deviance Table
## 
## Model 1: Nema ~ arsenic + cadmium + chromium + copper + iron + manganese + 
##     mercury + lead
## Model 2: Nema ~ arsenic + cadmium + chromium + copper + iron + manganese + 
##     lead
##   Resid. Df Resid. Dev Df  Deviance
## 1       101     2131.6             
## 2       102     2131.6 -1 -0.027794
Species richness
## FULL MODEL
## Adjusted R2 is: 0.16
Fitting linear model: S ~ arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 25 3.301 7.573 1.779e-11 * * *
arsenic 0.4272 0.6242 0.6845 0.4953
cadmium -88.04 42.61 -2.066 0.04139 *
chromium -0.02777 0.07971 -0.3484 0.7283
copper -0.2003 0.5601 -0.3576 0.7214
iron -9.051e-05 4.162e-05 -2.175 0.03197 *
manganese -0.0008686 0.002235 -0.3886 0.6984
mercury -14.76 55.56 -0.2656 0.7911
lead 1.086 1.169 0.9292 0.355
zinc 0.03835 0.211 0.1818 0.8561
Variance Inflation Factors
  arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 2.12 2.6 2.97 6.2 1.42 2.67 1.66 4.69 7.36 

## REDUCED MODEL
## Adjusted R2 is: 0.19
Fitting linear model: S ~ cadmium + iron + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 25.98 2.603 9.98 5.648e-17 * * *
cadmium -85.3 25.03 -3.408 0.0009242 * * *
iron -0.0001117 3.131e-05 -3.567 0.0005413 * * *
lead 0.7594 0.4005 1.896 0.06063
Variance Inflation Factors
  cadmium iron lead
VIF 1.56 1.09 1.63 

## Analysis of Variance Table
## 
## Model 1: S ~ arsenic + cadmium + chromium + copper + iron + manganese + 
##     mercury + lead + zinc
## Model 2: S ~ cadmium + iron + lead
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1    101 2307.2                           
## 2    107 2359.0 -6   -51.726 0.3774 0.8919

## RMSE for the full model: 4.989739
## RMSE for the reduced model: 4.761062
Total abundance
## FULL MODEL
## Adjusted R2 is: 0.07
Fitting linear model: N ~ arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 134.2 47.03 2.854 0.005261 * *
arsenic -2.777 8.863 -0.3133 0.7547
cadmium -1223 604.2 -2.024 0.04569 *
chromium 0.9859 1.123 0.8776 0.3823
copper -9.338 7.985 -1.17 0.245
iron -0.0005237 0.0005867 -0.8926 0.3742
manganese -0.04486 0.0315 -1.424 0.1576
mercury -879.8 785.1 -1.121 0.2652
lead 43.58 16.5 2.641 0.009615 * *
zinc 0.9465 3.013 0.3141 0.7541
Variance Inflation Factors
  arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 2.13 2.6 2.96 6.24 1.42 2.66 1.66 4.67 7.38 

## REDUCED MODEL
## Adjusted R2 is: 0.09
Fitting linear model: N ~ cadmium + copper + manganese + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 122.9 31.97 3.844 0.0002088 * * *
cadmium -714.2 364.8 -1.958 0.05293
copper -4.688 3.175 -1.476 0.1429
manganese -0.04996 0.01833 -2.725 0.007535 * *
lead 33.88 10.42 3.252 0.001545 * *
Variance Inflation Factors
  cadmium copper manganese lead
VIF 1.59 2.51 1.57 2.98 

## Analysis of Variance Table
## 
## Model 1: N ~ arsenic + cadmium + chromium + copper + iron + manganese + 
##     mercury + lead + zinc
## Model 2: N ~ cadmium + copper + manganese + lead
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1     99 449178                           
## 2    104 461228 -5    -12050 0.5312 0.7522

## RMSE for the full model: 73.60588
## RMSE for the reduced model: 69.0653
Shannon index
## FULL MODEL
## Adjusted R2 is: 0.14
Fitting linear model: H ~ arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.869 0.343 8.364 3.496e-13 * * *
arsenic 0.01789 0.06486 0.2759 0.7832
cadmium -2.401 4.428 -0.5422 0.5889
chromium -0.01677 0.008283 -2.024 0.04556 *
copper 0.004981 0.0582 0.0856 0.932
iron -5.231e-06 4.324e-06 -1.21 0.2292
manganese 0.0003217 0.0002323 1.385 0.1691
mercury 1.882 5.773 0.326 0.7451
lead -0.05074 0.1214 -0.4178 0.677
zinc 0.003013 0.02193 0.1374 0.891
Variance Inflation Factors
  arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 2.12 2.6 2.97 6.2 1.42 2.67 1.66 4.69 7.36 

## REDUCED MODEL
## Adjusted R2 is: 0.18
Fitting linear model: H ~ chromium + manganese
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.587 0.1636 15.81 7.149e-30 * * *
chromium -0.02062 0.004746 -4.343 3.179e-05 * * *
manganese 0.0002977 0.0001483 2.008 0.04719 *
Variance Inflation Factors
  chromium manganese
VIF 1.74 1.74 

## Analysis of Variance Table
## 
## Model 1: H ~ arsenic + cadmium + chromium + copper + iron + manganese + 
##     mercury + lead + zinc
## Model 2: H ~ chromium + manganese
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1    101 24.910                           
## 2    108 25.456 -7  -0.54588 0.3162 0.9452

## RMSE for the full model: 0.5485397
## RMSE for the reduced model: 0.4813905
Piélou’s evenness
## FULL MODEL
## Adjusted R2 is: 0.13
Fitting linear model: J ~ arsenic + cadmium + chromium + copper + iron + manganese + mercury + lead + zinc
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.7812 0.1128 6.924 4.137e-10 * * *
arsenic -0.02017 0.02133 -0.9456 0.3466
cadmium 2.122 1.456 1.457 0.1481
chromium -0.006769 0.002724 -2.485 0.0146 *
copper 0.001312 0.01914 0.06857 0.9455
iron 9.241e-07 1.422e-06 0.6498 0.5173
manganese 0.0001934 7.639e-05 2.531 0.01291 *
mercury 1.178 1.899 0.6203 0.5365
lead -0.05353 0.03994 -1.34 0.1832
zinc 0.001286 0.007211 0.1784 0.8588
Variance Inflation Factors
  arsenic cadmium chromium copper iron manganese mercury lead zinc
VIF 2.12 2.6 2.97 6.2 1.42 2.67 1.66 4.69 7.36 

## REDUCED MODEL
## Adjusted R2 is: 0.16
Fitting linear model: J ~ cadmium + chromium + manganese + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.8252 0.07737 10.67 1.757e-18 * * *
cadmium 1.949 0.9507 2.05 0.0428 *
chromium -0.00563 0.002017 -2.792 0.006219 * *
manganese 0.0001854 5.368e-05 3.454 0.0007948 * * *
lead -0.04977 0.0168 -2.962 0.00377 * *
Variance Inflation Factors
  cadmium chromium manganese lead
VIF 1.72 2.23 1.9 2 

## Analysis of Variance Table
## 
## Model 1: J ~ arsenic + cadmium + chromium + copper + iron + manganese + 
##     mercury + lead + zinc
## Model 2: J ~ cadmium + chromium + manganese + lead
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1    101 2.6946                           
## 2    106 2.7561 -5 -0.061575 0.4616  0.804

## RMSE for the full model: 0.1765139
## RMSE for the reduced model: 0.1614341

4. Multivariate regressions

These analysis have been done on PRIMER, with a DistLM to identify the variables that explain the most the community variability and with a dbRDA to plot the results.

All abiotic variables at BSI

2017 dbRDA Design 2

2017 dbRDA Design 2

Parameters

To be added!

Elliot Dreujou

6th February 2019